Chapter 5:

Market Risk

Nguyen Thi Lien

Faculty of Mathematical Economics, NEU
Email: [email protected]

4/27/2022 RISK ANALYTICS 1

What are we going to learn?
✓The VaR measure of risk, assumptions about return distributions and holding
period, and limitations of VaR.
✓Expected Shortfall (ES), compare and contrast VaR and ES
✓Define backtesting and exceptions and explain the importance of backtesting
VaR models.

4/27/2022 RISK ANALYTICS 2

 Estimating Returns
❑ The return losses is to quote return losses as positive values.
❑ Profit/loss data: P/Lt = Pt + Dt - Pt–1
Pt : the value of asset/portfolio at t and Dt: interim payments at t
❑ Estimating Returns:
Pt +Dt −Pt−1 Pt +Dt
◦ Arithmetic return data: 𝑟t = = −1
Pt−1 Pt−1

Pt +Dt
◦ Geometric return data: R t = ln
Pt−1

4/27/2022 RISK ANALYTICS 3

Exercise 5.1:
An investment has probabilities and giving one-year returns equal to:
Returns (R1) -10% 0% 10% 20% 30%
Prob 0.1 0.3 0.2 0.3 0.1
1. What is the mean return and the standard deviation of the return R1

2. Suppose that a portfolio P includes two investments (R1 and R2) with the equally
weightings. The mean return R2 of 8% and the standard deviation of 0.1%. What is
the total mean and standard deviation of returns if the correlation between them
is 0.2?

4/27/2022 RISK ANALYTICS 4

Value at Risk
VaR is a probabilistic method of measuring the potential loss in portfolio value
over a given time period and for a given distribution of historical returns.

Loss/profit

4/27/2022 RISK ANALYTICS 5

Example:
A firm has determined that the value at risk (VaR) of its investment portfolio is $18
million for one day at a 95% confidence level. Which of the following statements
regarding this VaR measure is correct?
A. There is a 95% probability that the portfolio will lose $18 million on a given day.
B. There is a 95% probability that the portfolio will lose no more than $18 million on a
given day.
C. There is a 5% probability that the portfolio will lose $18 million on a given day.
D. There is a 5% probability that the portfolio will lose no more than $18 million on a
given day

4/27/2022 RISK ANALYTICS 6

 Value at Risk
The VaR-measure is calculated as the quantile function of the loss distribution,
given a confidence level α, α ∈ [0, 1]. The expression of the VaR(α) is shown:
VaR(α) = inf{l ∈ (R) : P(L > l) ≤ 1 − α}
Example: VaR(10 days,99%) = $10 million
=> A VaR of USD 10 million would mean that we are 99% certain that the loss
during the next ten days will be less than USD 10 million.
or there is a probability of only 1% that the loss over the next ten days will be
greater than USD 10 million.
4/27/2022 RISK ANALYTICS 7
 Value at Risk
Historical-based approaches typically fall into three sub-categories:

- Parametric approach

- Nonparametric approach

- Hybrid approach

4/27/2022 RISK ANALYTICS 8

 Value at Risk - Parametric approach
If expected return other than zero is given:
VaR(1 days, α%) = µ + zα% .σ
Where:
zα% : the critical z-value based on the normal distribution.
µ, σ: the mean and standard deviation of daily returns on a percentage basis.
By the asset value:
VaR(α%) dollar = VaR(α%) decimal basis x asset value

4/27/2022 RISK ANALYTICS 9

Exercise 5.3: Value at Risk - Parametric approach
A risk management officer at a bank is interested in calculating the VaR of an
asset that he is considering adding to the bank’s portfolio. If the asset has a
daily mean of 1% and standard deviation of returns equal to 1.6% and the
asset has a current value of $5 million, calculate the VaR at 5% significance.

4/27/2022 RISK ANALYTICS 10

 Value at Risk - Nonparametric approach
The simplest way to estimate VaR is by means of historical simulation (HS),
estimated by means of ordered loss observations

4/27/2022 RISK ANALYTICS 11

 Value at Risk - Nonparametric approach
Historical Simulation Approach

Calculating and estimating VaR with a historical simulation approach:

(1) order return observations from largest to smallest.

(2) The observation that follows the threshold loss level denotes the VaR limit.

(3) More generally, the observation that determines VaR for n observations at the
(1 - α) confidence level would be: (α × n) + 1

4/27/2022 RISK ANALYTICS 12

 Value at Risk - Nonparametric approach
Historical Simulation Approach
Example: we have 1000 loss observations and are interested in the VaR at the 95%
confidence level.
➢ The confidence level implies a 5% tail => there are 50th observations in the tail
➢ We can take the VaR to be the 51st highest loss observation

4/27/2022 RISK ANALYTICS 13

Exercise 5.6: Historical VaR
You have accumulated 100 daily returns for your $100,000,000 portfolio. After
ranking the returns from highest to lowest, you identify the lowest six returns:

-0.0011, -0.0019, -0.0025, -0.0034, -0.0096, -0.0101

Calculate daily value at risk (VaR) at 5% significance using the historical method.

4/27/2022 RISK ANALYTICS 14

Exercise 5.3: Historical VaR
Calculate VaR of the 5th percentile Lowest Historical
Rank
Returns simulation weigh
using historical simulation and the
1 −4.70% 0.01
data provided:
2 −4.10% 0.01
3 −3.70% 0.01
4 −3.60% 0.01
5 −3.40% 0.01
6 −3.20% 0.01

4/27/2022 RISK ANALYTICS 15

 Value at Risk - Nonparametric approach
VaR can be converted from a 1-day basis to a longer basis by multiplying
VaR by square root of the number of days (T) in the longer time period
(called the square root rule)

VaR(X%) T_days = VaR(X%) 1-day x 𝑻

Or VaR can converting to different confidence levels:

𝒁𝟏%
VaR(1%) = VaR(5%) x
𝒁𝟓%

4/27/2022 RISK ANALYTICS 16

Value at Risk - Nonparametric approach
Monte Carlo Simulation
❑ The computer softwares generates hundreds, thousands, or even millions of
possible outcomes from the distributions of inputs specified by the user.
❑ The several thousand weighted average portfolio returns will naturally form a
distribution, which will approximate the normal distribution.
❑ Using the portfolio expected return and the standard deviation, which are
part of the Monte Carlo output.

4/27/2022 RISK ANALYTICS 17

Value at Risk - Hybrid approach
Weighted Historical Simulation
Uses historical simulation to estimate the percentiles of the return and weights
that decline exponentially. The following three steps are required:
Step 1: Assign weights for historical realized returns to the most recent n
(1−𝜆)𝜆𝑖−1
returns using: 𝑤 𝑖 =
1−𝜆𝑛
Step 2: Order the returns.
Step 3: Determine the VaR for the portfolio by starting with the lowest return
and accumulating the weights until X percentage is reached.
4/27/2022 RISK ANALYTICS 18
 Exercise 5.7: Weighted Historical Simulation
Calculate the initial VaR at the 5% using a hybrid approach with the most recent 100
observations (n = 100) and λ= 0.96.
Hybrid
Lowest Number of Hybrid cummulative
Rank returns past periods weight weight
1 -4.7% 2 0.0391 0.0391
2 -4.1% 5 0.0346 0.0736
3 -3.7% 55 0.0045 0.0781
4 -3.6% 25 0.0153 0.0934
5 -3.4% 14 0.0239 0.1173
6 -3.2% 7 0.0318 0.1492
4/27/2022 RISK ANALYTICS 19
Exercise 5.7: Weighted Historical Simulation
The initial VaR at the 5% using a hybrid approach with the most recent 100
observations (n = 100) and λ= 0.96.
Hybrid
Lowest Number of Hybrid cummulative
Rank returns past periods weight weight
1 -4.7% 2 0.0391 0.0391
2 -4.1% 5 0.0346 0.0736
3 -3.7% 55 0.0045 0.0781
4 -3.6% 25 0.0153 0.0934
5 -3.4% 14 0.0239 0.1173
6 -3.2% 7 0.0318 0.1492
4/27/2022 RISK ANALYTICS 20
 Expected Shortfall
Expected shortfall (ES) is a risk measure that does take account of expected losses
beyond the VaR level. Expected shortfall (which is also called condition VaR or tail
loss) is the expected loss conditional that the loss is greater than the VaR level.
ESα = E(L|L ≥ VaRα)
Another definition that often is used is:
1 1
ESα = ‫׬‬𝛼
VaR 𝑢 (L)du
𝛼

It can be interpreted as that expected shortfall is calculated by averaging VaR over

all levels u ≤ α
4/27/2022 RISK ANALYTICS 21
Exercise 5.4:
An investment has a uniform distribution where all outcomes between -40
and +60 are equally likely. What are the VaR and expected shortfall with a
confidence level of 95%?

4/27/2022 RISK ANALYTICS 22

Exercise 5.6:
A one-year investment has a 2% chance of losing USD 9 million, 9% chance
of losing USD 4 million, and a 89% chance of gaining USD 1 million.
1. What are the VaR and the expected shortfall when the confidence level is
95% and the time horizon is one year?
2. Suppose that there are two independent identical investments
mentioned. What are the VaR and the expected shortfall for a portfolio
consisting of the two investments when the confidence level is 95% and the
time horizon is one year?

4/27/2022 RISK ANALYTICS 23

 Expected Shortfall
When losses are normally distributed with mean µ and standard deviation σ,
the expected shortfall is
−𝑧2ൗ
𝑒 2
ESα = μ + 𝜎
(1−𝑋) 2𝜋

where X is the confidence level,

z is the critical value in the standard normal distribution that has a probability
X% of being exceeded.

4/27/2022 RISK ANALYTICS 24

Exercise 5.5:
The distribution of the losses from a project over one year has a normal
loss distribution with a mean of -10 and a standard deviation of 20.
1. what is the one-year VaR when the confidence level is 95%, 99% and
99.9%
2. The change in the value of the project over two years? Assume that
changes in successive years are independent.
3. The one-year expected shortfall with a confidence level of 95%, 99%
and 99.9%?

4/27/2022 RISK ANALYTICS 25

Exercise 5.5: Calculating ES - Historical Simulation
You collected your bank's trading book's daily mark-to-market profit & loss
(P&L) for the last two years, which is 500 trading days. The ten worst losses
were (in millions, losses as positives): 18.0, 15.0, 15.0, 13.0, 9.0, 9.0, 8.0, 7.0,
6.0, 5.0. What is the historical 99.0% confident daily expected shortfall (ES;
aka, conditional VaR or expected tail loss)?
a. $9.0 million
b. $10.5 million
c. $14.0 million
d. $16.5 million
4/27/2022 RISK ANALYTICS 26
 Exercise 5.4: Historical VaR
You collected the portfolio’s daily profit and loss (P&L) for the most recent 100
traded days. You sort the outcomes form worst to best. Below are the ten worst
days.
Rank 1 2 3 4 5 6 7 8 9 10
Loss -11.0 -10.0 -9.0 -7.0 -3.0 -2.0 -2.0 -1.5 -1.3 -1.0
You want to the estimate the portfolio’s VaR under the HS approach, and its
expected shortfall. Which are, respectively, the portfolio’s 95% VaR and ES?
a, 95% VaR is either {3.0 ;5.0 or 7.0} and 95% ES is 3.0
b,95% VaR is either {2.0 ;3.0 or 2.5} and 95% ES is 8.0
c, 95% VaR is either {2.0 or 3.0} and 95% ES is either {7.0 or 8.0}
d, 95% VaR is 3.0 and 95% ES is either {7.0 or 8.0 or 9.3}
4/27/2022 RISK ANALYTICS 27
Backtesting VaR Models
❑ Backtesting is the process of comparing losses predicted by a value at risk (VaR)
model to those actually experienced over the testing period.
❑ The number of actual observations that fall outside a given confidence level are
called exceptions. The number of exceptions falling outside of the VaR confidence
level should not exceed (1 - confidence level).
❑ For example:
Exceptions should occur less than 5% of the time if the confidence level is 95%.

4/27/2022 RISK ANALYTICS 28

 Backtesting VaR Models
The importance of backtesting:
❑ The tool for providing model validation to ensure that actual losses do not exceed
expected losses at a given confidence level.
❑ The risk managers and regulators are properly calibrated or accurate. If the level of
exceptions is too high, models should be recalibrated and risk managers should
reevaluate assumptions, parameters, and/or modeling processes.
❑ The Basel Committee allows banks to use internal VaR models and backtesting to test
the adequacy of those internal VaR models.
❑ Bank regulators rely on backtesting to verify risk models

4/27/2022 RISK ANALYTICS 29

Model Backtesting with Exceptions
❑The probability of exception:
𝒑 = 𝟏 − 𝒄𝒐𝒏𝒇𝒊𝒅𝒆𝒏𝒄𝒆 𝒍𝒆𝒗𝒆𝒍 = 𝟏 − 𝐜
❑ Failure rates define the percentage of times the VaR confidence level is exceeded
in a given sample.
𝑵
Failure rate = 𝟏 −
𝑻
Where N represents the number of exceptions and T represents the sample size
❑ Failure rate is unbiased if the computed 𝑝 approaches the confidence level as the
sample size increases.

4/27/2022 RISK ANALYTICS 30

Exercise 5.11.
In backtesting a value at risk (VaR) model that was constructed using a 97.5%
confidence level over a 252-day period.

1. What is an acceptable probability of exception for exceeding this VaR amount?

2. How many exceptions are forecasted?

4/27/2022 RISK ANALYTICS 31

Exercise 5.11.
1. We expect to have exceptions 2.5% of the time (1 − 97.5%).
◦ If exceptions are occurring with greater frequency, we may be underestimating the
actual risk.
◦ If exceptions are occurring less frequently, we may be overestimating risk and
misallocating capital as a result.
-> Non-parametric tests can then be used to see if the number of times a VaR model
fails is acceptable or not.
2. The number of exceptions = 1 − 0.975 × 252 = 6.3

4/27/2022 RISK ANALYTICS 32

Basel Committee Rules for Backtesting
Basel Penalty Number of capital multiplier
Increase in k
Zones Exceptions (k)
Green 0 to 4 3.00 0.00
Yellow 5 3.40 0.4
6 3.50 0.5
7 3.65 0.65
8 3.75 0.75
9 3.85 0.85
Red 10 or more 4.00 1.00

4/27/2022 RISK ANALYTICS 33